Menü
When and how to write scenario holdouts for factory verification.
Mit Codex oder Claude installieren Kopieren Sie diesen Prompt, fügen Sie ihn in Codex, Claude oder einen anderen Assistant ein und lassen Sie die Skill-Seite prüfen und installieren.
Basierend auf der SOC-Berufsklassifikation
| name | scenario-authoring-guide |
| description | When and how to write scenario holdouts for factory verification. |
| for | human |
Write scenarios BEFORE or ALONGSIDE specs — never after implementation. Scenarios written after implementation are no longer holdouts; the builder has already seen the context that shaped them.
2–5 scenarios per project. Cover major user paths without becoming a test suite. More scenarios add cost without adding coverage signal.
Write from the user's perspective, not the developer's. Describe what the user does and what they observe — not what functions are called or what files change.
Focus on the non-obvious: edge cases, error recovery, interaction patterns the spec might not make explicit. The spec covers the happy path; scenarios stress-test the assumptions underneath it.
Scenarios MUST NOT be shared with Builder — holdout by design. If Builder sees
scenarios, they become implementation targets rather than independent verification
criteria. The scenarios/ directory is Builder-inaccessible for this reason.
For factory infrastructure work, evaluate against scenarios/meta/factory-itself.md
directionally (does this change move the factory closer to the meta-scenario?),
not as a pass/fail gate.
Complete reference for the factory CLI — all commands, arguments, options, and usage. Use when an agent needs to invoke factory commands, check correct syntax, or understand available options.
When and how to write a human-action-needed entry to memory/{agent}/needs.md. Agents write these entries when they hit blockers that only a human can resolve.
When you encounter a failure, correction, or new pattern, write a learning and optionally propose a skill.
How the factory workspace is organized and what each directory means.
What to examine during a reflection pass and how to write observations to memory/{agent}/needs.md.
How to integrate scenario evaluation into verification reports.
